remove irda usage of isa_virt_to_bus()
Quoted 10 lines · William Lee Irwin III
* Removed uses of
isa_virt_to_bus
This resolves Debian BTS #218878.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=218878From: Paavo Hartikainen <[email protected]>
To: Debian Bug Tracking System <[email protected]>
Subject: IrDA modules fail to load
Message-Id: <[email protected]>
Quoted 5 lines
When trying to "modprobe irda irtty", it fails with following message:
FATAL: Error inserting irda (/lib/modules/2.6.0-test9/kernel/net/irda/irda.ko): \
Unknown symbol in module, or unknown parameter (see dmesg) And in "dmesg" I see \
these: irda: Unknown symbolisa_virt_to_bus
irda: Unknown symbolisa_virt_to_bus
Could you please send this directly to me. I hate scrubbing
large patches from the mailing list archive.
Note that before even thinking of pushing this patch in the
kernel, we need to perform testing with the hardware on i386 and potentially on ARM. The author only tried with irtty that doesn't use this function, so that's not a valid test at all. Finding people test those changes is going to be tough, as usual.
I'm also wondering about the validity of those changes, but
that's another matter I need to go through. During 2.5.X, some people assured me that using isa_virt_to_bus was safe on all platform with an ISA bus...
Thanks... Jean
Quoted 11 lines · Jean Tourrilhes
Could you please send this directly to me. I hate scrubbing
large patches from the mailing list archive.
Note that before even thinking of pushing this patch in the
kernel, we need to perform testing with the hardware on i386 and potentially on ARM. The author only tried with irtty that doesn't use this function, so that's not a valid test at all. Finding people test those changes is going to be tough, as usual.
I'm also wondering about the validity of those changes, but
that's another matter I need to go through. During 2.5.X, some people assured me that usingisa_virt_to_buswas safe on all platform with an ISA bus...
Okay, well, I myself didn't produce this, and I couldn't tell offhand if it was bogus or not. I presumed bugreporter made happy and spraying it across the debian userbase was enough to verify it at runtime. From what you're telling me, this is not the case.
Can you recommend people to do this kind of testing?
Apparently people aren't entirely happy with "dump it on lkml and wait for an ack or nak", which I've noted for future reference, but probably won't have a need to consider again (it's very rare that I have to deal with changes I didn't write myself or are otherwise in areas I don't have much knowledge about). OTOH, it was easier to find than buried in a distro BTS and/or cvs, not that that makes it ideal.
-- wli
Quoted 13 lines · William Lee Irwin III
Quoted 11 lines · Jean Tourrilhes
Could you please send this directly to me. I hate scrubbing
large patches from the mailing list archive.
Note that before even thinking of pushing this patch in the
kernel, we need to perform testing with the hardware on i386 and potentially on ARM. The author only tried with irtty that doesn't use this function, so that's not a valid test at all. Finding people test those changes is going to be tough, as usual.
I'm also wondering about the validity of those changes, but
that's another matter I need to go through. During 2.5.X, some people assured me that usingisa_virt_to_buswas safe on all platform with an ISA bus...
Quoted 2 lines
Okay, well, I myself didn't produce this, and I couldn't tell offhand if it was bogus or not.
I can't either, that's why we need to check it.
Quoted 3 lines
I presumed bugreporter made happy and spraying
it across the debian userbase was enough to verify it at runtime. From what you're telling me, this is not the case.
IrDA is a special case because there are few users and most
are still using 2.4.X.
Quoted 1 line
Can you recommend people to do this kind of testing?
Me. I'll also post the patch to the linux-irda mailing list.
Quoted 5 lines
Apparently people aren't entirely happy with "dump it on lkml and wait for an ack or nak", which I've noted for future reference, but probably won't have a need to consider again (it's very rare that I have to deal with changes I didn't write myself or are otherwise in areas I don't have much knowledge about).
I'm not a full time Linux hacker, I follow lkml from the
archive. No big deal.
Quoted 2 lines
OTOH, it was easier to find than buried in
a distro BTS and/or cvs, not that that makes it ideal.
Debian is usually very good sending me bug reports (especiall
on wireless tools), so I'm a bit surprised that it did not work this time. But I've seen a recent trend by Debian to do more Debian specific stuff for system level config, which I find disturbing.
Quoted 1 line
-- wli
Thanks !
Jean
Quoted 2 lines · William Lee Irwin III
Quoted 2 lines
OTOH, it was easier to find than buried in
a distro BTS and/or cvs, not that that makes it ideal.
Quoted 4 lines · Jean Tourrilhes
Debian is usually very good sending me bug reports (especiall
on wireless tools), so I'm a bit surprised that it did not work this time. But I've seen a recent trend by Debian to do more Debian specific stuff for system level config, which I find disturbing.
These seem to be the ones that have leaked through the cracks. Apart from this (and the posting is helping to find all the concerned parties) things should go largely through normal channels in general.
Also, thanks for doing the footwork to get this thing properly evaluated, as it was unclear to me the risks involved. I don't want bogus code going in.
-- wli